摘要
文章在复合Linex对称损失下,就参数的先验分布为伽玛分布,共轭先验及无信息先验进行了对比,得到艾拉姆咖分布参数的唯一贝叶斯估计。进一步通过R软件进行随机模拟,得到参数的Bayes估计值,从而得出结论:当先验为共轭先验时,估计值较接近艾拉姆咖分布参数的真值。
Under compound Linex symmetric loss, this paper compares different prior distributions of the parameter as gamma distribution, conjugate prior and uninformative prior to obtain the unique Bayesian estimation of Эрланга distribution parameter. Furthermore, the paper uses R software for stochastic simulation and derives the Bayesian estimate of the parameter, thereby coming to a conclusion that when the prior is conjugate prior, the estimate value is relatively more proximate to the true value of Эрланга distribution parameter.
引文
[1]Varian.A Bayes Approach to Real Estate Assessment[A].Fienberg SE,ZELLENER A.Studies in Bayesian Economics and Statistics inHonour of Lenoard J.Savage[C].Amesterdam:North Holland,1975.
[2]张睿.复合Linex损失下的参数估计[D].大连:大连理工大学硕士论文,2007.
[3]龙兵.不同先验分布下艾拉姆咖分布参数的Bayes估计[J].数学的实践与认识,2015,(4).
[4]龙兵.不同损失函数下艾拉姆咖分布参数的Bayes估计——全样本情形[J].重庆师范大学学报:自然科学版,2013,(5).